The abundance of vascular endothelial growth factor (VEGF) receptors on surface of tumor vessel endothelial cells (ECs) is an attractive target for therapy. We propose to use a genetically engineered form of VEGF-A (VE0F121) labeled for radiotherapy of primary tumors in mice as well as for SPECT imaging for the serial non-invasive assessment of VEOF receptor density. This new protein combines the two 3-112 aa fragments of VEGFI2I (normally a dimer) fused into a single-chain (sc) with a N-terminal l5-aa tag containing a unique(C4) cysteine residue for site-specific labeling with simple, facile and non-destructive maleimide-based chemistry without byproducts or unmodified protein;problems that are inherent with other labeling techniques. scVEGF also has no pro-angiogenic domain thereby eliminating the risk of increasing tumor angiogenesis and the spread of metastases. Efforts to starve tumors of blood supply by targeting ECs have had mixed success and may be due to the unintended drug selection of resistant hypoxic tumor cell clones. We propose a new strategy which leverages the ability of scVEGF linked to isotopes to target tumor vessels and kill malignant cells directly. The most appropriate isotopes appear to be those with intense beta (Lu-177 or Y-90) or Auger electron emissions (In-Ill) that can kill cells within a small local range while sparing adjacent tissues. In this proposal we will model the size, geometry, location (with respect to tumor margins), and VEGF receptor density of the angiogenic rim over time in tumor bearing mice using biodistribution, autoradiographic of Lu-I77 (or Y-90) labeled scVEGF and immunohistochemical assays. These studies will be conducted with several well characterized cell lines;411 syngeneic mammary, human MDA mammary and the human HT29 colonic carcinoma models. We will study both small (2-3 mm) and large tumors (6-9 mm) in all lines and with our lucifcrase expressing 41'I and MDA constructs we will also examine tumors that are non-palpable but can he seen with BLI imaging (very small <1mm) to ascertain the relationships of the angiogenic rim to tumor size over time. From these data we will then perform dosimetry calculations and screen normal mice for signs ofradiotoxicity at the chosen tumor dose prior to actual radiotherapy. Short term, long-term (survival), and dose escalation radiotherapy studies will then be performed and compared with untreated (-) controls and Sutent (anti-angiogenic) treatment (+ controls). Successful completion of our proposal will give the needed preclinical data for pursuing scVEGF targeted radiotherapy in other models and ultimately clinical applications.